China Plans To Bankrupt The AI Industry In The United States
What To Do When The Competition Gives Away The Product
Then there are the neocloud companies like Coreweave, Nebius and Iren building out the data centers, leveraging piles of capital and credit, creating expensive and huge buildings across the country.
Not to be forgotten are the chip builders, mainly Nvidia, along with Advanced Micro Devices and Intel.
All these companies are spending all or most of their days working to make better, faster, more impressive AI models come to life.
While we could look at a number of problems with AI (like not enough paying customers, their sometimes unreliable generated outcomes, their ethical structure whether regarding students doing their homework or how the AI companies acquired all that data in the first place) right now let's look at an entirely different problem.
Let's focus on the external problem of China.
If we look at China's tie to the United States economy it's been a dynamic story over the last 50 years. US companies started sending some production over to China in the 1980s but it was very simple manufacturing, be it brooms, toys, very basic products. By the 1990s the manufacturers in China were getting better, they started making more complex products like small motors, kitchen parts, more B2B inputs. Skip to 2026 and the situation has grown out of hand. Now China is the leading manufacturer for just about anything you can name, including cars, robots, batteries, renewable energy, just about all the inputs that go into everything in the US.
China provides more than 90% of the rare earths, essential to making most high-end products. China makes more than 40% of the manufactured inputs that go into the equipment used by the US Military.
Now we come to AI. China is swinging at the AI future with two fists. With one hand they are creating generative AI platforms that are a fraction of the cost of the United States AI firms are charging. With the other hand they are working to create lower cost high-end microchips, to someday match or surpass what TSMC (Taiwan) and Nvidia are able to design.
The first problem is the greater difficulty. The Mag Seven, the neoclouds, the US chip companies have combined to spend over $1.5 trillion on generative AI since 2022, and that number is growing quickly, the rate of growth is increasing. Some estimates are that spending on AI in the United States will exceed $800 billion in 2026, and the projections for 2027 are substantially larger.
All this spending is going to need an eventual payday. All this investment, all the loans and issued bonds are going to need to be repaid. As we quickly race past the $2 trillion figure of investment, realize that collectively these tech companies are going to have to be invoicing a lot of people, companies, government customers at a hard and fast clip to get that money back for the investors. And by the way, these are the companies with the highest market capitalizations in the stock market, all the stock market gains over the last three years can be attributed to the massive infusion of cash and capital expenditure made for AI.
Guess what? The leaders in China can see this just as easily as we can. They've probably quantified the figures far more than we have. They know what a precarious situation the US business community has put itself in here. The expectation by 99% of business leaders is that AI will just keep growing at a fast clip, that all sorts of customers will emerge to pay for these services on a monthly basis, and the journey of massive spending will just continue forward.
What if China sees that pending, highly needed outcome from their perch on the other side of the world? It is absolutely essential that someone or some entity start paying much more money on these AI services than they have up to this point. The United States has over 4,000 data centers up and running, with many more in the pipeline.
China sees this danger and they have unleashed dozens of nimble AI companies that are working to provide American business and consumers with the same level of generative AI services, at a much lower price. That lower price is running about 90% off the US market price (using Anthropic and OpenAI pricing as the benchmark) - so let's call it what it is, China is giving away their AI services.
Imagine a Chevrolet truck offered for $60K and then a product from China arrived in the US market with a competitive truck for sale, and there was no detectable difference between the Chevy and the China truck, but the Chinese version was only $6,000? Remember, nothing would be visually or functionally different between the two trucks, no quality difference, perceived or real, between the two. On the sales lot you would see a $60K truck for sale and a $6K truck for sale. What's going to happen in the market? Obviously the business for the Chevrolet version is going to quickly diminish, if not vanish.
We can cut from that theoretical story to a real event that just happened on June 16, 2026 where Microsoft announced they were considering using a Chinese AI provider call DeepSeek to offer a cheaper version of their Microsoft Copilot AI tool to "cut enterprise AI costs" - and they stressed it was just for the cheap version of Copilot, no need to worry, the important AI models will still be using the good old USA produced AI for that.
Sure. Just like how in 1989 we (the US) used China to 'just' make our brooms and toys and simple plastic manufactured items. We know how this story ends. Companies are addicted to finding the lowest cost option, or in another sense, they have to. If Microsoft starts using low cost Chinese AI services then what will Meta or Alphabet do, keep using the far more expensive US services exclusively? They will get priced out in the marketplace if they do, so it'll just be a matter of time until they also switch to a China-based AI provider for their low cost services.
We could go into further nuance here. China is producing an AI product that is open weight, meaning that the end user can manipulate the functionality of that AI, it's an open system that can be far more manipulated that the main closed weighted versions that OpenAI and Anthropic provide. Basically, the Chinese versions are similar to open source software where the US versions are proprietary software. We could talk about potential government intervention, perhaps the US will not allow US companies to use the Chinese versions. We could talk about the potential hacking and information grift that the China versions might secretly be working on behind the scenes.
We could discuss all the different China companies in this field, DeepSeek, Moonshot, Alibaba, Zhipu (now Z.ai) or others, but this is not meaningful either. Just like the character X from the movie JFK, when discussing the assassination he said: "Does it really matter who shot from what rooftop?" - it's the same here, all the debates about laws, privacy, safety, quality, ethics - in the end a super cheap product that works will win, no matter what regulations or hesitations are in the way.
Once the cat is out of the bag, the details are not very meaningful. It gets back to the $60K truck or the $6K truck. Everyone is going to want a $6K truck that runs with high quality and proven efficiency. Period. All the other discussions you will hear in the coming months and years are background noise.
The reality is that China has come out with quality AI service at really low prices. There is a very good chance those low prices will only be heading lower as more and varied AI models keep emerging out of the now strutting and supremely confident China. The trend toward open weighted, low cost, Chinese models is now an inevitable force. Even on the chance the United States puts a ban on using these cheap AI companies, that won't stop all the other nations of the world from using them. The US will just be left behind, on a North American island, while the rest of the world runs fast and far ahead of American business.
Speaking of trucks, we can see this same situation playing out now with cars. As US car manufacturers and the US Government defend the domestic automobile market, Chinese car companies like BYD are running away with the market in Asia, Europe, Africa and South America. Chinese cars are even selling into Canada and Mexico, despite tariffs and quotas.
China-made AI will travel the same path as their car manufacturers, a rapid and dynamic success story.
With that assumption in place, we (the United States) now face quite a dilemma. While the water's surface of 2026 looks pretty serene, if you go down just below the surface it is anything but calm. There's a problem so big it's a bit hard to fathom and make the mind understand. The United States has a huge stake on their AI being the hot in-demand product around the world. If China begins overtaking the US, if they begin taking any real market share, that will cast a sudden shadow across a very dreamy vision.
The number of ramifications are stunning to consider.
The stock market is at all-time highs, mostly on the backs of the AI and tech complex. These unrealized gains in the S&P 500, thousands of ETFs and mutual funds, are in the trillions of dollars. The total value of the US stock market is over $75 trillion, while the US economy is only $32 trillion. Sounds like some lofty expectations have been built into the future, much of which is that the AI complex in America will create amazing business outcomes, huge profits, massive growth. Those might be the things that China is planning on stopping.
The top 20% of Americans (by wealth) own 87% of that stock market. And that same top 20% of Americans accounts for 60% of consumer spending each year in the United States. We have built a dangerous line of dominoes with our AI super obsession. US-based AI has pushed the stock market to spectacular levels, which is turn has made the top echelon of Americans extremely wealthy, which in turn has turned them into consumer machines.
What if China's push to provide extremely cheap (almost free) AI destroys the future earnings expectations for US tech companies? Do the wheels fall off of the stock market? And if that happens, do we work backwards where Americans who felt super rich now feel much humbler when it comes time to spend money on travel, restaurants, products? The potential here for a negative feedback loop is scary because of what it could mean not so much for the stock market itself, but for what that could do to the overall economy.
Then there the potential financial losses themselves, not from the markets, but from the AI industry itself. If China keeps successfully pushing AI services and they take market share, and in the process force US providers of AI to charge less money to compete, then where does that leave the now approaching $2 trillion in investment? The investors who bought into AI (the equity investors) want to make a profit, but perhaps their investment will go down instead of making a return. And then there's the lenders, what about the firms that issued loans to the various tech firms? Seems like if China gives the service away it will make it difficult to make money with US-based AI, which in turn will make it really difficult to pay back the borrowed funds.
The overall financial hit to the US economy could be truly staggering if China runs away with even half of the US market in AI. The losses are hard to project but would certainly be in the trillions.
Finally, there is another loss that will come from China clobbering the AI industry in the US and that's prestige. Americans tend to assume that we will be first in the world in so many ways, militarily, technologically with innovation and creativity, entrepreneurship, financial markets, but do we (and the rest of the world) still believe that way if China runs away with our AI tech the same way they ran away with manufacturing vacuum cleaners? We probably do not and that will be an odd spot to land in, because it'll be a post World War II moment that is unlike anything we've seen since 1945. Since that moment we've been assumed to be the leaders in human intelligence, constantly holding the technology edge over everyone else in the world. And that will not be true any longer once we lose the war for AI.